In recent years the availability of automated DNA synthesizers has revolutionized the fields of molecular biology and biochemistry. As a result, linear DNA oligonucleotides of specific sequences are available commercially from several companies. These can be used for a variety of applications. For example, DNA oligonucleotides can be used as primers for cDNA synthesis, as primers for the polymerase chain reaction (PCR), as templates for RNA transcription, as linkers for plasmid construction, and as hybridization probes for research and diagnostics.
DNA and RNA oligonucleotides, i.e., oligomers, also can act as sequence-specific inhibitors of gene expression through binding of a complementary, or "antisense," base sequence. See, for example, E. Uhlmann et al., Chem. Rev., 90, 543 (1990), and Oligodeoxynucleotides: Antisense Inhibitors of Gene Expression; J. S. Cohen, Ed.; CRC Press: Boca Raton, Fla., 1989. These antisense oligomers have been shown to bind to messenger RNA at specific sites and inhibit the translation of the RNA into protein, splicing of mRNA or reverse transcription of viral RNA and other processing of mRNA or viral RNA. In addition, "anti-gene" oligomers have been developed that bind to duplex DNA and inhibit transcription.
Strong inhibitory activity has been demonstrated in vitro and in vivo using these antisense and anti-gene oligomers against viruses such as HIV-1, Herpes Simplex Virus, and influenza virus, among others, as well as against several types of cancer. Thus, antisense and anti-gene oligonucleotides could be used as antiviral and anticancer agents and therapeutic agents against almost any disease mediated by gene expression. In addition, in some cases improved activity has been reported for analogs of DNA, including DNA and RNA phosphorothioates and 2'-O-methyl ribonucleotides. All potential therapeutic applications, however, would require large amounts (tens or hundreds of grams) of specific oligomers for animal and clinical trials, and even more for eventual use as a pharmaceutical. See, for example, I. Kitajima et al., Science, 258, 1792 (1992), and M. Z. Ratajczak et al., PNAS, 89, 11823 (1992).
Ribozymes are naturally occurring RNA sequences that possess the property of self-catalyzed (autolytic) cleavage. Known ribozymes include, for example, hairpin and hammerhead motifs. The catalytic "hammerhead" domain of a ribozyme typically contains 11-13 conserved nucleotides at the juncture of three helices precisely positioned with respect to the cleavage site. A hammerhead containing less than 60 contiguous nucleotides was found to be sufficient for rapid autolytic cleavage in the absence of any protein (D. E. Ruffner et al., Gene, 82, 31-41 (1989)).
Autolytic cleavage by ribozymes is an intramolecular event and is referred to as occurring "in cis". However, the essential constituents for a biologically active RNA hammerhead structure can be present on separate molecules. For example, one strand may serve as a catalyst and the other as a substrate. Ribozymes acting in trans can, for example, interfere with the production of a protein by cleaving the target mRNA transcript encoding the protein. Reddy et al. disclosed ribozyme cleavage of a target RNA in trans using a synthetic RNA molecule containing a hammerhead (catalytic region) flanked by sequences designed to hybridize to the target RNA substrate on either side of the potential cleavage site. (U.S. Pat. No. 5,246,921, issued Sep. 21, 1993). The novel ribozyme was capable of selectively cleaving the bcr-abl mRNA of a cell containing the Philadelphia Chromosome, thereby blocking synthesis of the BCR-ABL protein associated with some forms of leukemia.
One major drawback in the use of oligonucleotides as diagnostic tools or therapeutic agents is the high cost of oligonucleotide synthesis by machine using the standard solid-phase synthetic methods. Reasons for this include the high costs of the synthetically modified monomers, e.g., phosphoramidite monomers, and the fact that up to a tenfold excess of monomer is used at each step of the synthesis, with the excess being discarded. Costs of DNA oligonucleotides have been estimated at $2-5 per base for one micromole (about 3 mg of a 10 mer) on the wholesale level. On this basis, 1 gram of a 20-base oligomer would cost on the order of $20,000. Thus, significant in vivo testing of antisense oligomers will be quite expensive until ways are found to lower the cost.
Enzymatic methods have the potential for lowering the cost of oligonucleotide synthesis. Enzymatic methods use DNA or RNA nucleotide triphosphates (dNTPs or NTPs) derived from natural sources as the building blocks. These are readily available, and are less expensive to produce than phosphoramidite monomers. Generally, this is because the synthesis of the nucleotide triphosphates from base monophosphates requires as little as one step. See, for example, E. S. Simon et al., J. Org. Chem., 55, 1834 (1990). Nucleotide triphosphates (NTPs) can also be prepared enzymatically. In addition, the polymerase enzymes used in these methods are efficient catalysts, and are also readily available.
There are two major methods now in use for enzymatic amplification of DNA: cloning and the polymerase chain reaction (PCR). See, for example, J. Sambrook et al., Molecular Cloning; 2nd ed.; Cold Spring Harbor Press, 1989, and R. K. Saiki et al., Science, 239, 487 (1988). Cloning requires the insertion of a double-stranded version of the desired sequence into a plasmid followed by transformation of a bacterium, growth, plasmid re-isolation, and excising the desired DNA by restriction endonucleases. This method is not feasible for large-scale preparation because most of the material produced (the vector) is in the form of unusable DNA sequences. PCR is a newer technique that uses a thermostable polymerase to copy duplex sequences using primers complementary to the DNA. Subsequent heating and cooling cycles allow efficient amplification of the original sequence. For short oligomers, such as those used in anti-sense applications (e.g., less than about 50 nucleotides), PCR is inefficient and not cost-effective because it requires a primer for every new strand being synthesized.
Recently, a method was developed for the enzymatic synthesis of DNA oligomers using a noncleavable linear hairpin-shaped template/primer in a PCR-like enzymatic synthesis. See G. T. Walker et al., PNAS, 89, 392 (1992). Although this method may be more cost-effective than PCR, the polymerase must still dissociate from the template to enable amplification. Furthermore, the end groups of the DNA produced are ragged and not well defined.
Other methods of DNA replication are discussed in Harshey et al., Proc. Nat'l. Acad. Sci., USA, 78, 1090 (1985); and Watson, Molecular Biology of the Gene (3rd Edition). Harshey et al., discuss the theoretical method of "roll-in" replication of double-stranded, large, circular DNA. The "roll-in" process involves small, double-stranded circle cleavage and incorporation into a genome. It is primarily a process for inserting double-stranded plasmids into a double-stranded genome. Although one could conceivably use an entire genome to replicate an oligonucleotide, the resulting product would be thousands of nucleotides longer than desired. Thus, the "roll-in" process would be a very inefficient means to produce target oligonucleotide sequences. Watson briefly mentions the replication of single-stranded circles, but the author focuses the reference on the replication of double-stranded circles.
Prior to the present invention, it was thought by those skilled in the art that processive rolling-circle synthesis would not proceed without additional proteins which unwind the duplex ahead of the polymerase. See, e.g. Eisenberg et al., PNAS USA, 73:3151 (1976); The Single-Stranded DNA Phages, D. T. Denhardt et al., eds., Cold Spring Harbor Press; Cold Spring Harbor (1978); and DNA Replication, W. H. Freeman, San Francisco, 1980. In Eisenberg et al., the in vitro replication of .phi.X174 DNA using purified proteins is disclosed. Among the listed necessary proteins are DNA unwinding protein (also known as SSB, single-strand binding protein), cisA protein, and rep protein. These DNA unwinding proteins (which require ATP) are necessary for this replicative synthesis; otherwise the polymerase stalls. The Single-Stranded DNA Phases includes a discussion of the mechanism of replication of a single-stranded phage and furthermore shows a scheme for this replication in FIG. 8 therein. One of the beginning stages of replication involves the elongation of a single-stranded (-) template annealed to a full-length linear (+) strand. Any further elongation necessarily requires unwinding of the helix ahead of the polymerase. DBP (Double-strand binding protein) was thought to be necessary to coat the displaced strand in order for there to be successful DNA synthesis during elongation.
The polymerase from phage .phi.29 is known to amplify DNA strands as large as 70 kb in length. Even though this polymerase exhibits such a high degree of processivity, the use of the polymerase from phage .phi.29 still results in the wasteful (in both time and monetary resources) production of unwanted nucleotides. In order to replicate an oligonucleotide prior to the present invention, those of skill in the art would have encoded the oligonucleotide as only a small portion of the entire replicated region. Moreover, utilizing a plasmid or phage method to replicate an oligonucleotide would require the investigator to first separate the strands and then purify the oligonucleotide from thousands of other base pairs.
RNA oligomers are currently synthesized by two principal methods: DNA synthesizer and enzymatic runoff transcription. Methods have been recently published for the use of a synthesizer to construct RNA oligomers using a modification of the phosphoramidite approach. See, for example, S. A. Scaringe et al., Nucleic Acids Res., 18, 5433 (1990). Chemical synthesis of RNAs has the advantage of allowing the incorporation of nonnatural nucleosides, but the yield decreases significantly as the length of the RNA product increases; stepwise yields of only 97.5% per round of synthesis are typical. Moreover, because of the need for additional protecting groups, RNA phosphoramidite monomers are considerably more expensive than are the DNA phosphoramidite monomers, making RNA synthesis by this method extremely costly. An alternative, the enzymatic runoff transcription method, utilizes a single or double-stranded DNA template. Runoff transcription requires a phage polymerase promoter, thus a DNA strand .about.20 nucleotides longer than the RNA desired must be synthesized. There are also strong sequence preferences for the RNA 5' end (J. F. Milligan et al., Nucleic Acids Res., 15, 8783-8798 (1987)). In runoff transcription the RNA copy begins to form on the template after the phage polymerase promoter and runs until the end of the template is reached. This method has the disadvantages of producing RNA oligomers with ragged, ill-defined end groups and giving relatively slow amplification. Both chemical synthesis and runoff transcription produce a number of undesired products shorter or longer than the desired RNA, lowering effective yields and requiring careful purification.
Double-stranded DNA plasmid vectors can be constructed to encode ribozymes as well as their self-cleavage sites, leading to self-processing after transcription (A. M Dzianott et al., Proc. Natl. Acad. Sci. USA, 86, 4823-4827 (1989); C. A. Grosshans et al., Nucleic Acids Res., 19, 3875-3880 (1991); K. Taira et al., Nucleic Acids Res., 19, 5125-5130 (1991)). However, plasmid vectors contain thousands of nucleotides extraneous to those required for the actual desired transcript, making them highly inefficient templates for ribozyme synthesis; shorter sequences generally possess greater activity. Their large size also poses problems for delivery into cells in cases where transcription is to be performed intracellularly. Also, plasmid vectors require promoters to initiate transcription.
Thus, there is a need for a low-cost, fast, and efficient method for the production of DNA and RNA oligomers having well-defined ends on a large scale. In view of their great therapeutic potential, new methods for in vitro and in vivo synthesis of ribozymes are particularly needed. In addition, there is a need to produce DNA and RNA analogs, such as, for example, DNA phosphorothioates, RNA phosphorothioates, and 2'-O-methyl ribonucleotides, with well-defined ends on a large scale and in an efficient manner. Furthermore, there is a need for a method that uses readily available enzymes and a readily prepared template to generate large amounts of a complementary sequence.